Monitor 001  /  Check-in 08 of 8  /  SEPTEMBER 2026

SEPTEMBER 2026Is the 2028 crisis happening?

Two analysts wrote a memo from June 2028 about an AI-driven crash. Every month we check which parts of it are coming true. This is the September 2026 reading, kept exactly as it was written. Source memo: THE 2028 GLOBAL INTELLIGENCE CRISIS, “What follows is a scenario, not a prediction.”.

So far · Machines: Happening. People: Not yet. Money: Other way.

Summary

An unusually useful month: one important part of the thesis got substantially stronger while another was directly contradicted by the data. Private-credit stress worsened to a new record, and Wipro gave the cleanest real-world example yet of AI producing enormous labour-productivity gains. But software, the sector the memo's credit mechanism depends on, has the lowest default rate of any major sector, and Wipro redeployed its freed capacity rather than firing it. The evidence now supports a coherent alternative version of the story rather than the memo's own.

Status of each link

MACHINESAre the machines good enough, and are companies using them?
PEOPLEAre people losing work and income because of it?
MONEYIs it breaking anything in the financial system?

New evidence in September 2026

2026-09-10Wipro says AI has freed capacity equivalent to 20,000 workers. It laid off none of them.

About 8% of a 243,000-person workforce. CTO Sandhya Arun told Reuters the staff were redeployed to other projects, retrained, or assigned to work alongside agents; more than 100,000 employees have advanced AI training. 'It doesn't necessarily mean person-to-person replacement by an agent.'

One event, two verdicts. The memo's version: AI productivity → redundant workers → layoffs → income loss → consumption decline. Wipro's version: AI productivity → redundant capacity → redeployment → fewer future hires and different jobs. Supports the technological premise, weakens the assumed short-term transmission into unemployment.

Touches 02 (supports), 03 (contradicts) · Reuters — Wipro's AI push frees capacity equivalent to 20,000 workers, CTO says

2026-09-15Software private-credit defaults fall to 0.6%, the lowest of any sector, while the aggregate sets a record 6.3%

Fitch, trailing twelve months to August, about 1,300 borrowers. Software down from 1.2% in July and 2.0% a year ago. Healthcare 9.9%, industrial manufacturing 9.9%, consumer products 8.7%. Fourteen default events in August against three in July; 89 unique defaulters; interest deferrals and PIK are about 47% of default events.

We now have enough data to distinguish 'private credit is stressed' from 'AI is causing private-credit stress'. The first is increasingly true. The second currently is not.

Touches 11 (contradicts), 10 (supports) · Fitch Ratings, via BigGo Finance — US Private Credit Default Rate Climbs to Record 6.3%; Software Falls to 0.6%

2026-09-04August payrolls +162,000; unemployment 4.1%; participation up to 61.6%; unemployment among bachelor's holders 2.7%

Information lost 23,000 jobs after averaging about 8,000 a month of losses; computing infrastructure and data processing lost 8,000, publishing 7,000. Restaurants (+59,000) and local-government education (+42,000) were about 62% of the whole gain. Construction added 22,000 on data-centre demand.

Service, manual and public-sector hiring healthy; the information and knowledge economy noticeably weaker. Closer to the memo's distributional thesis than the 4.1% headline suggests, and still not an aggregate white-collar event.

Touches 04 (contradicts), 05 (contradicts) · Bureau of Labor Statistics — The Employment Situation, August 2026

2026-09-01July JOLTS: openings 7.27m, hires 5.1m, quits 3.1m, layoffs 1.7m at a 1.0% rate

An economy entering the memo's displacement spiral should show rising involuntary separations. It is not. Employers are adjusting through slower hiring, attrition, redeployment and selective restructuring.

Touches 04 (contradicts), 03 (contradicts) · Bureau of Labor Statistics — Job Openings and Labor Turnover, July 2026

2026-08-28July nominal consumer spending +0.2%, real PCE essentially flat; real disposable income +0.4%; saving rate 3.0%

Softer than June's +0.4% real reading. Nowhere near the demand vacuum the memo requires, but the next few reports matter more because the savings cushion is thin.

Cooling, not collapsing.

Touches 06 (unresolved) · Bureau of Economic Analysis — Personal Income and Outlays, July 2026

2026-09-11Oracle adds $700m to expected restructuring costs, taking the FY26 programme to about $2.8bn; raising about $40bn in debt and equity; free cash flow minus $5.4bn

Job reductions and contract terminations alongside the AI and cloud expansion.

Reinforces the alternative scenario: the financial accident may come because companies borrow and spend enormous amounts on AI infrastructure whose returns prove insufficient.

Touches 14 (supports), 03 (supports) · Reuters — Oracle shares rise as AI cloud backlog beats estimates

2026-09-10OpenAI launches a financial-services product built with Morgan Stanley and Evercore: research, modelling and pitchbooks, connected to LSEG, PitchBook and Daloopa

Precisely the high-wage knowledge work whose automation matters to the thesis, far more than chatbot customer service.

Workflow compression, not yet investment bankers disappearing at macro scale.

Touches 02 (supports), 03 (unresolved) · Reuters — OpenAI launches ChatGPT for the financial services industry

2026-09-10Visa, Mastercard and Ant International announce a common framework for identifying and verifying AI agents making purchases

Agentic commerce is happening. The incumbents are becoming the authentication and trust layer for agents.

A fairly strong counterexample to the memo's idea that agents route around today's payment intermediaries.

Touches 09 (contradicts) · Reuters — Payment firms Visa, Mastercard, Ant International team up on AI agent trust framework

2026-09-17The software ETF that fell 21% in the first quarter is back to flat on the year

IGV year-to-date total return of 0.04% as of 17 September. The SaaSpocalypse round-tripped.

Touches 09 (contradicts) · Stock Analysis / iShares — iShares Expanded Tech-Software Sector ETF (IGV)

2026-09-02San Francisco home prices up 6% year on year; Seattle down 3.6%

Redfin attributes San Francisco to concentrated AI wealth and Seattle to layoffs at Amazon, Microsoft, Meta and Expedia. The memo forecast San Francisco down 11%, Seattle down 9%, Austin down 8%.

The memo treated tech metros as one bloc. They are moving in opposite directions, and the AI capital is the strong one.

Touches 12 (contradicts) · Redfin — A Tale of Two Tech Cities

Check-in

Surprise

  • Wipro produced the output of 20,000 additional workers and cut nobody.
  • Software private-credit defaults halved in the year software was supposed to break the credit market.
  • The AI infrastructure boom is creating construction, electrical and manufacturing jobs, a mechanism the memo underweights.

Still waiting for

  • Material deterioration in white-collar household credit, concentrated in tech metros rather than spread across all borrowers
  • Any large PE-backed software default, the memo's Zendesk moment
  • Labour's share of GDP turning down
  • A single quarter in which a card network reports slowing consumer spending
  • Software private-credit defaults climbing from 0.6% while white-collar hiring keeps weakening, which would finally connect two separate pieces of the chain

Previous question resolved

Asked in March 2026Will software defaults follow the aggregate private-credit defaults up?No, and it is now emphatic. Software fell to 0.6% while the aggregate rose to a record 6.3%. The two moved apart, not together, for seven straight months.

Previous question resolved

Asked in August 2026If AI-attributed layoffs are falling while adoption rises, is redeployment now the default corporate response?On the evidence of the largest single case this year, yes. Wipro redeployed rather than dismissed, and framed it as an operating model rather than a cost event.

Previous question resolved

Asked in August 2026Whether Proofpoint and Medallia are the first of many or two outliersTwo outliers, so far. The sector rate halved in the month after they surfaced.

What to watch next

  • Information-sector employment
  • Real consumer spending against a 3.0% saving rate
  • Software private-credit defaults, hardest of all
  • Whether Wipro's redeployment holds when the next capacity release arrives
  • Q3 sector attribution from Fitch and Proskauer

Research note

The evidence supports a coherent alternative story: AI gets extremely productive, companies need fewer incremental white-collar workers, hiring falls and existing workers are redeployed, knowledge-sector bargaining power deteriorates, gains accrue to firms and capital, and enormous infrastructure spending creates jobs elsewhere. That can produce substantial inequality, sectoral disruption and credit losses without the memo's immediate depression-like spiral. The three numbers to watch hardest into October: information-sector employment, real consumer spending, and software private-credit defaults. If that last one climbs from 0.6% while white-collar hiring keeps weakening, two currently separate pieces of the chain would finally connect.

← August 2026   All check-ins

A predicted symptom appearing does not mean the predicted mechanism caused it.